A known ion detector for a mass spectrometer comprises a microchannel plate (“MCP”) detector. A microchannel plate consists of a two-dimensional periodic array of very small diameter glass capillaries (channels) fused together and sliced into a thin plate. The microchannel plate detector may comprise several million channels, each channel operating in effect as an independent electron multiplier. An ion entering a channel will interact with the wall of the channel causing secondary electrons to be released from the wall of the channel. The secondary electrons are then accelerated towards an output surface of the microchannel plate by an electric field which is maintained across the length of the microchannel plate by applying a voltage difference across the microchannel plate.
The secondary electrons generated by an incident ion will travel along a channel on parabolic trajectories until the secondary electrons strike the wall of the channel and cause further secondary electrons to be generated or released. This process of generating secondary electrons is repeated along the length of the channel such that a cascade of several thousand secondary electrons may result from the incidence of a single ion. The secondary electrons then emerge from the output surface of the microchannel plate and are detected.
It is known to provide two microchannel plates sandwiched together and operated in series. The two microchannel plates are maintained at a high gain so that a single ion arriving at the first microchannel plate may cause a pulse of, for example, 107 or more electrons to be emitted from the output surface of the rearmost of the two microchannel plates. The two microchannel plates may be arranged in a chevron arrangement wherein the microchannel plates are arranged in face to face contact such that the channels in one microchannel plate are arranged at an angle with respect to the channels of the other microchannel plate. This arrangement helps to suppress ion feedback which may otherwise lead to damage.
The requirements of an electron multiplier in a Time of Flight mass spectrometer are particularly stringent. The electron multiplier should produce minimal spectral peak broadening and provide a linear response at both low and high ion arrival rates whilst allowing single ion events to be distinguished clearly from electronic noise.
In order to achieve these criteria the output of an electron multiplier due to an individual ion arrival event should have minimal temporal spread and the pulse height distribution of the electrons should be as narrow as possible. In addition, the gain of the electron multiplier should preferably be in the order of 106 or greater to allow single ion events to be easily distinguished from electronic noise.
For ion counting applications microchannel plate ion detectors have so far yielded the most satisfactory characteristics in terms of these criteria. However, under optimal operating conditions the dynamic range of microchannel plate ion detectors can be limited.
Under conditions of high gain, for example 106-107, the output current from a single channel of a microchannel plate will become space-charge saturated, leading to narrow pulse height distributions approaching gaussian distributions. Narrow pulse height distributions are advantageous for ion counting devices using Time to Digital Converters (“TDC”) as they allow the majority of single ion events to be distinguished from electronic noise. Narrow pulse height distributions are also advantageous for use with Analogue to Digital Converters (“ADC”) as they allow for accurate quantitation at low count rates and an improved dynamic range.
The maximum output current of a microchannel plate detector is limited by the recovery time of the individual channels after illumination and the total number of channels illuminated per unit time. Ions incident upon a microchannel plate detector in an orthogonal acceleration Time of Flight mass analyser will illuminate a discrete area of the microchannel plate detector. Accordingly, ions will be incident upon only a portion of the total number of microchannels available regardless of the area of the microchannel plate. Therefore, when large ion currents are incident upon the microchannel plate ion detector or at certain steady state output currents a significant proportion of channels will not recover fully after illumination and hence the overall gain of the microchannel plate ion detector will be reduced. In particular, the final 20% of the length of the channels in the final gain stage of a microchannel plate ion detector will be limited by this saturation point first. This has the result of causing there to be a non-linearity in the response of the ion detector for quantitative analysis which will result in inaccurate isotopic ratio determinations and inaccurate mass measurements.
In order to increase the maximum input event rate which the ion detector can accommodate before saturation occurs, the gain of the microchannel plate could in theory be reduced. However, reducing the gain would cause broadening of the pulse height distribution and would shift the pulse height distribution to a lower intensity resulting in a compromise in the ability of the ion detector to detect all single ion arrivals above the threshold of electronic noise.
The limitations of a conventional microchannel plate ion detector will now be considered in more detail below. In particular, two microchannel plates arranged as a chevron pair will be considered. After a cloud of electrons has exited an individual channel in a microchannel plate the charge within the channel walls must be replenished. For a circular microchannel plate the number of channels N is given by:   N  =            π      ⁢                           ⁢              D        2                            12            ⁢              p        2            where D is the diameter of the microchannel plate and p is the channel centre to centre spacing (channel pitch).
For a circular microchannel plate having a diameter of 25 mm and comprising channels having a diameter of 10 μm and a channel pitch of 12 μm, the total number of channels N is 3.9×106. Typically, the total resistance of such a single microchannel plate is 108 Ω. Therefore, the resistance Rc of a single channel of the microchannel plate is approximately 3.9×1014 Ω.
The total capacitance of a single microchannel plate may be approximated by considering it to be a pair of parallel metal plates separated by a relatively thin glass plate. The total capacitance C may be approximated as:            ⁢      C    =                  ɛ        ⁢                                   ⁢                  ɛ          0                ⁢        S            d      where C is the capacitance in Farads, ∈ is the dielectric of glass (approximately 8.3 F/m), ∈0 is the permittivity of a vacuum 8.854×10−12, S is the area of the microchannel plate and d is the thickness of the microchannel plate.
Therefore, if the thickness d of the microchannel plate is taken to be 0.46 mm, the total capacitance C of a single microchannel plate is 78 pF and hence the capacitance Cc for each channel of the microchannel plate is 2×10−17 F.
The time constant τ for recovery of an individual channel in the microchannel plate after an ion event is given by:CcRc=τ
In this example the time constant τ for an individual channel is 7.8 ms. For a pair of microchannel plates in a chevron pair arrangement a primary ion event at the input surface of the first microchannel plate typically results in secondary electrons illuminating approximately ten channels on the input surface of the second microchannel plate. Assuming the first and second microchannel plates are identical, then the maximum ion input event rate E at the first microchannel plate is given by:   E  =      N          10      ⁢                           ⁢      τ      
Accordingly, the maximum ion input event rate Emax at the first microchannel plate which is sustainable without appreciable overall loss of gain of the whole ion detector is approximately:       E    max    =      E    10  
In the example given above the maximum input event rate Emax is 5×106 events/s. At a mean gain of 5×106 this equates to a maximum output current Imax of 4×10−6 A.
Orthogonal acceleration Time of Flight mass spectrometers commonly have very large ion currents at sampling repetition rates of tens of kHz. Under these conditions the input ion current to the microchannel plate approximates to a steady DC input current. The gain of the microchannel plate is constant until the microchannel plate output current exceeds approximately 10% of the available current passing through the microchannel plate, i.e. strip current. In the example given above the maximum output current Imax is 10−6 A when 1000 V is maintained across the microchannel plate.
Several approaches have been developed to overcome this limitation in the maximum output current from a microchannel plate. For example, reducing the resistance of the microchannel plate reduces the time constant τ for channel recovery and increases the strip current available and hence increases the maximum output current from the microchannel plate. However, there are also practical limitations. The negative temperature coefficient of resistance of the channel walls in the microchannel plate ultimately results in thermal instability as the resistance of the microchannel plate is reduced. This causes heating of the microchannel plate which can result in ion feedback leading to thermal runaway which may result in local melting of the microchannel plate glass. The mechanism by which heat is dissipated from a microchannel plate is predominantly by radiation from the surface of the microchannel plate and the heat dissipation is therefore directly proportional to the exposed surface area of the microchannel plate.
It has been found experimentally that it is not practical to operate microchannel plates at levels of heat generation above 0.01 W/cm2. For a circular microchannel plate having a diameter of 33 mm and maintained at a bias voltage of 1000 V, this rate of heat generation corresponds to a microchannel plate having a total resistance of approximately 107 Ω. As a consequence of this limitation on the microchannel plate total resistance, it should be noted that the maximum output current of the microchannel plate cannot be increased by simply decreasing the diameter of the channels in the microchannel plate in order to increase the number of channels available per unit area. For example, a circular microchannel plate having a diameter of 33 mm, corresponding to an active diameter of 25 mm, and comprising channels having a diameter of 10 μm and a channel pitch of 12 μm will have a total of 3.9×106 channels. If the microchannel plate has a total resistance of 107 Ω then the resistance of each channel will be 3.9×1013 Ω. For a circular microchannel plate having the same diameter, the same total resistance, a reduced channel diameter of 5 μm and a reduced channel pitch of 6 μm the total number of channels will be 1.6×107. Accordingly, each channel will now have an increased resistance of 1.6×1014 Ω. In this example, it is shown that by reducing the diameter and pitch of the channels in the microchannel plate the total number of channels has increased by a factor of approximately ×4. However, the resistance per channel and hence the time constant for recovery of an individual channel τ has also increased by the same factor. Therefore, no overall gain in the maximum output current of the microchannel plate is obtained.
Direct cooling of the microchannel plate does in theory allow very low resistance microchannel plates to be employed. However, such direct cooling is impractical in most situations.
Another method of increasing the maximum output current of the microchannel plate is to disperse the incoming ion beam over a relatively large microchannel plate or over the input surface of multiple microchannel plates. This dispersion of the ion beam increases the number of channels available without changing the characteristics of the individual channels in the microchannel plate. The overall resistance of the microchannel plate ion detector is therefore reduced resulting in a higher available strip current and hence a higher onset level of channel saturation.
In this arrangement the microchannel plate(s) may be operated under relatively stable conditions since the surface area available for radiative cooling of the microchannel plate(s) is also increased. However, deliberately diverging the ion beam as it travels towards the ion detector is impractical in many situations depending on the geometry and size of an individual mass spectrometer. Furthermore, in order to diverge the ion beam electric fields must be provided in the region of the mass spectrometer upstream of the ion detector. This is particularly disadvantageous in a Time of Flight mass spectrometer in which the region upstream of the ion detector is a drift region since the introduction of an electric field into the drift region may affect the resolution and mass measurement accuracy of the ion detection system. In addition, the electric field conditions are required to be changed when detecting negative and positive ions. Therefore, diverging the ion beam is not a practical solution to this problem.
It is therefore desired to provide an improved detector for a mass spectrometer.